Realtime object extraction and tracking with an active camera using image mosaics

Abstract

Moving object extraction plays a key role in applications such as object-based videoconference, surveillance, and so on. The difficulties of moving object segmentation lie in the fact that physical objects are normally not homogeneous with respect to low-level features and it's usually tough to segment them accurately and efficiently. Object segmentation based on prestored background information has proved to be effective and efficient in several applications such as videophone, video conferencing, and surveillance, etc. The previous works, however, were mainly concentrated on object segmentation with a static camera and in a stationary background. In this paper, we propose a robust and fast segmentation algorithm and a reliable tracking strategy without knowing the shape of the object in advance. The proposed system can real-time extract the foreground from the background and track the moving object with an active (pan-tilt) camera such that the moving object always stays around the center of images.